The Rhetoric and the Reality of the Legal Guarantee for Freedom … · 2020. 2. 4. · control –...
Transcript of The Rhetoric and the Reality of the Legal Guarantee for Freedom … · 2020. 2. 4. · control –...
「정부행정」 제10권(2014): 103-148Journal of Government Administration, Vol.10 (2014)
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information:
A Cross-National Study on the Level of Freedom of Information
Nam, Taewoo**1)
Abstract
Analyzing the longitudinal data combining several global indices, this paper
empirically examines what factors determine the degree to which a country achieves
freedom-of-information ideals – access to information, press freedom, and
transparency. Though countries with a freedom-of-information law show greater
achievement of freedom-of-information ideals, the presence or absence of the
legislation itself is not significant for estimating accomplishment of the ideals after
controlling both time-series and cross-sectional effects. Such determinants as the
general quality of public administration and socio-political stability proved significance
as an explanatory variable in the panel data regressions. Countries with Scandinavian
and German legal traditions are higher achievers of freedom-of-information ideals than
those with English and French heritages. From a narrow focus on the subsample of
countries with a freedom-of-information law, the study found that not merely the
intention behind a particular provision but also concreteness and specificity of the
statute significantly affect achievement of the legal ideals.
Key Words: Freedom of information; Access to information; Right to know;
Transparency; Openness; Freedom of press
Ⅰ. Introduction
A law pursues to actualize our desires for a better society. The
legislation for freedom of information (FOI) is a specific expression of
legal protection for the right to know. We expect the law to make a
* Assistant Professor of the Department of Public Administration at Myongji University
104 「정부행정」 제10권(2014)
government more accountable and transparent via a broader avenue to
approaching governmental information, to which citizens and the press are
more easily accessible. In this sense, I identify FOI ideals as the public’s
access to information, freedom of press and transparency.
Amongst about 210 countries on our globe, only 73 countries have the
law that legally stipulates freedom of information. This plain fact leads to
skepticism of global FOI. We should say that if an FOI law (FOIL) itself
is an effective tool to realize FOI ideals, merely one third of all countries
promise freedom of information. However, the level of the effectiveness of
the law, despite plausibility of the general tendency that countries with an
FOIL are high achievers of FOI ideals, varies from country to country.
Furthermore, not all countries with an FOIL do achieve the legal ideals
successfully. If an FOIL per se is actually symbolic and mythical in spite
of its legal guarantee for FOI ideals, it has no better than nominal values.
With such skeptical consideration of FOI realities in a global context, I
raise a simple research question involving the rhetoric and the reality of
an FOIL: What determines the realization of FOI ideals in global
societies?
I hypothesize that the general quality of public administration,
socio-political stability and legal tradition serve as a more important
determinant for the degree of fulfilling FOI ideals than the simplistic fact
that an individual country has or does not have an FOIL. Focusing on
countries that have an FOIL, I expect the difference in FOIL features
would make a distinction between high achievers and low achievers of
FOI ideals. For a cross-national study, this paper, employing the panel
data regressions and analysis of variance (ANOVA) as main
methodologies, statistically analyzes the longitudinal data generated by
merging several worldwide indices from 2001 to 2008. The paper starts
with reviewing core arguments in previous studies. After expatiating on the
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 105
panel data and variables employed, its main parts report key findings and
implications from statistical analyses. The final sections address limitations
of this cross-national empirical study and conclusive remarks.
Ⅱ. Ideals and Realities of FOI Laws
Since Sweden’s pioneering adoption of an FOIL in 1776, there has been
a long temporal lag until a global spread of the legislation during recent
decades. With a tidal wave of democratization over the globe, the impetus
factors like anti-corruption movement, international pressures, further
modernization and societal informatization have triggered consideration of
an FOIL as the necessity for democracy (Banisar, 2006). The law states
the right of access to information (ATI) in order to make government
accountable. It can be effective in three levels. In the basic level, it
prevents a government from interfering with information dissemination. Its
middle level requires a government to comply with citizens’ demands on
governmental information. On the more mature stage of an FOIL regime,
we expect a government’s affirmative obligation to inform public. Each
level of these hierarchical goals may be either a desired state for some
countries or a current reality for others. The gulf between ideals and
realities differentiates a group of greater FOI countries from the other
group of less FOI ones.
1. FOI ideals: Right-to-know, press freedom and transparency
FOI is regarded as a basic human right (Perritt & Lhulier, 1997;
Piotrowski & Rosenbloom, 2002). The fundamental purpose of an FOIL is
106 「정부행정」 제10권(2014)
to give the public the right to know and define a legal framework for the
sharing of information (Islam, 2006). The law has provided “an engine
that makes publicly available much of the vast and otherwise inaccessible
storehouse of government information” (Gellman, 1997). The right to
access to government-held information is a cornerstone of representative
democracy (Mendel, 2008). As with the First Amendment to the United
States Constitution, the Freedom of Information Act (FOIA) purports to
“promote democratic-constitutional values including transparency and
individual rights” (Piotrowski & Rosenbloom, 2002).
An FOIL is an essential tool to ensure democratic control of
government. Feinberg (2004) identified the FOIA as “a logical extension of
decades-long efforts to hold government agencies accountable for their
decisions.” Without the right of access to information, a citizen’s ability to
participate actively in policy deliberations or hold public institutions
accountable for their conduct is compromised and even constrained. By
signifying reduction in restrictions imposed on information flow (Islam,
2006), an FOIL keeps people aware of policymaking processes (North,
1999) and the value of public information (Chongkittavorn, 2002). In the
United States, the FOIA has been practiced to increase governmental
transparency and advance accountability (Piotrowski & Rosenbloom, 2002
). The British expectation for the legislation goes toward eroding cultural
and institutional secrecy by making the government more open (Worthy,
2008).
Main users and beneficiaries of an FOIL are the public and the media.
Improved access to official records has not been granted only to an
amorphous populace but to the press as the most influential champion for
the law (Relyea, 1979). An FOIL holds little value for citizenry without
free media (Martin & Feldman, 1998); thus, free press is a precondition
to enhance citizens’ right to know. A FOIL’s legal protection of the right
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 107
to know helps the press perform its reportorial and journalistic work of
documenting governmental activities in public interests (Rush, 1986).
2. Determinants for FOI realities: Administration, stability, legal heritage, and provisions
FOI statutes are so widely diffused that they have come to resemble
“rationalized institutional myths” (Meyer & Rowan, 1991). Aside from
practices of implementing an FOIL, the broad provisions provide myths
and symbols to demonstrate that a country legally supports good values
– i.e. freedom of information, transparency, openness and democratic
control – for public administration. The legal rhetoric is similar across
countries with an FOIL. The FOIA of the United States has been a
template for other countries that prepare for a new FOIL, and thus most
FOILs around the world are much alike under the influence of such a
model law (Mendel, 2008). However, the interpretation and application of
an even identical provision may vary with country, culture, regime and
polity because of differences in administrative quality and disparities in
social, cultural and legal contexts. This divergence can be a reason for the
dissonance between our desires for FOI ideals and the current reality.
Having an FOIL is not enough to ensure that it is effective (Foerstel,
1999; Islam, 2006; Roberts, 2000). The nominal existence of the law
cannot be any hallmark for its successful implementation. Government
agencies must be required to publish information, and there must be some
mechanisms to implement the law effectively. Gellman (1997) identified
problems inherited in the FOIA as three facets: 1) poorly drafted, 2)
inadequately funded, and 3) unenthusiastically implemented. As the undue
level of flexibility and administrative discretion is problematic for
implementation, manifold problems with the FOIA are administrative
108 「정부행정」 제10권(2014)
rather than legislative.
In Commonwealth countries with longer histories of FOI legislation, the
creation of a FOI regime has not ended secrecy within a government
(Roberts, 2005). When considering the historical fact that the FOIA in the
United States was legislated in the face of enormous opposition by the
executive branch, less effective and less committed implementation of the
law has no wonder. Though today’s climate is not as hostile as the
executive branch in the earlier FOIA regime was (Relyea, 2009), public
agencies seldom have rigid motives for faithful administration of the
statutes. Lombard’s (2007) experiment on 408 local public bodies’
responses to information requests at the state of Illinois reported that an
average citizen encounters blatant resistance, numerous delays and arbitrary
obstructions, which violate the original spirit of the FOIA and the state
FOIL.
Regarding that information is a fragile, time-sensitive commodity
(Feinberg, 1986), a delayed and inconsistent response is a frequent
complaint of the press who needs the information in time for a deadline
(Rush, 1986). Worthy (2008) found sources of inconsistency and tardiness
in the low quality of administration and information management.
Inextricably entwined in a larger context of information policy, an FOIL is
an object of management and implementation apart from its normative
ideals (Feinberg, 1986). The degree to which the FOI ideals are fulfilled
depends on administrative discretion to implement the laws (Darch &
Underwood, 2005; Davis, 2000; Feinberg, 2004; Halstuck, 2000; Islam,
2006). A key to realization of FOI goals is the change in attitude within
agencies in regard to FOIA administration (Relyea, 1979).
Another problem that public administration faces in implementing an
FOIL arises from the ongoing mantra of public entrepreneurialism. Since
the emergence of the New Public Management (NPM) paradigm, public
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 109
administrators and managers have rejected a conventional stress on
processes and democratic-constitutional values in favor of results
(quantifiable performance measures). The businesslike management trend
categorizes process-oriented actions to implement an FOIL as nonessential
and noncore missions, especially in terms of budget priority (Piotrowski &
Rosenbloom, 2002; Roberts, 2000). The reinvention movement may
undermine laws that give citizens the right of access to government
information.
In addition to the quality of public administration, the FOI reality is
not independent of socio-political conditions and a legal infrastructure.
Political stability affects the degree to which the rule of a law is
effectively and seamlessly implemented and enforced. In the reverse way, a
national FOIL represents a crucial opportunity to consolidate stability of a
country (Mistry, 2006). Accordingly, the level of domestic socio-political
stability has a positive influence on the accomplishment of FOI ideals.
Legal tradition is a rudimental background to determine the degree to
which FOI ideals are realized. The influence between law and society is
reciprocal. While a law affects societies and culture, a country’s legal
framework evolves with diverse social, cultural factors. Legal tradition may
change over time due to information exchange across different cultures
and societies. A case law acts as a driver of internal information
exchange; meanwhile, globalization is an external catalyst for variability in
legal tradition. Institutionalized legal heritage, nevertheless, reflects a solid
socio-cultural context that a country has maintained for long. Most
countries largely fall into several branches of legal tradition: common law
tradition (American or British) and civil law tradition (Roman, German or
French). Although specific provisions in a national FOIL have a great
resemblance across countries that lately benchmarked some influential
models of FOI legislation in the Western world, the reference to disparate
110 「정부행정」 제10권(2014)
legal traditions will distinguish the pattern in interpretation and application
of a similar FOIL.
The last determinant for FOI reality is a specific feature of an FOIL.
What items the law is armed with can be an important predictor for
explaining the level of FOI ideals accomplishment. FOILs vary in scope
(the degree of specificity and applicability) from country to country. Some
laws are very detailed regarding what information is kept secret under
what circumstances, and others are quite general (Martin & Feldman,
1998). As Gellman (1997) pointed out, the reason why FOI ideals are not
achieved may be a problem inherent in provisions. Along with other
factors, statute specificity will determine the level of FOI reality.
Ⅲ. Methods and Measurements
Based on the discussion of FOI ideals and realities, I raise a research
question: “What determines the extent to which ideals of an FOI law are
realized?” FOI ideals that this study focuses on include access to
information (ATI) as an individual citizen’s right to know, freedom of the
press, and a government’s transparency. As a set of explanatory variables
to estimate the level of FOI realities, the general quality of public
administration, socio-political stability, legal tradition and FOIL features
will present an answer to the research question.
1. Dataset
This study constructs a new panel dataset by merging multiple
secondary longitudinal datasets that consist of various yearly country-level
indicators. Analyzing panel data instead of cross-sectional data allows us
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 111
to know the impact of focal explanatory variables on dependents while
controlling for year-specific and country-specific noisy influences.
Additionally, the increased number of observations will reduce the
possibility of violating the normality assumption and mitigate a potential
problem from the lack of data in particular years and units of the original
datasets.
<Table 1> The number of countries included in each index
Dependent variables Independent variables
Freedom House 1
TransparencyInternational(CorruptionPerceptionIndex) 2
GlobalIntegrityIndex 3
Worldwide Governance Indicators 4
Worldwide Governance Indicators 4
from(Porta et al.,
2008; Djankov et al., 2003) 5
from(Tromp, 2008) 6
YearPress
freedomTransparency ATI Public Admin Stability Legal FOIL’s
2001 186 90 – – – 188 2162002 193 102 – 197 190 188 2162003 193 132 – 197 200 188 2162004 194 145 25 204 207 188 2162005 194 155 – 204 208 188 2162006 194 163 41 206 209 188 2162007 194 180 50 207 209 188 2162008 – 180 46 – – 188 216Total 1,348 1,147 162 1,215 1,223 1,504 1,728
Source 1: http://www.freedomhouse.org/template.cfm?page=1Source 2: http://www.transparency.org/policy_research/surveys_indices/cpiSource 3: http://www.globalintegrity.org/data/downloads.cfmSource 4: http://www.govindicators.orgSource 5: http://www.economics.harvard.edu/faculty/shleifer/paperSource 6: http://www3.telus.net/index100/foi
As shown in Table 1, both dependent variables and independent
variables have different data sources. The indicator of press freedom is
acquired from Freedom House online. Transparency International annually
provides the indicator of national corruption. This index has been broadly
used as a proxy for national transparency in previous cross-national
studies. The open dataset that Global Integrity releases presents surveys
pertinent to a citizen’s access to government information in limited number
112 「정부행정」 제10권(2014)
of countries (Appendix 3). I derived the quality of public administration
and political stability from World Bank’s Worldwide Governance
Indicators (WGI). This study imported the taxonomy of legal heritages
categorized in the political-economics papers of Porta, Lopez-de-Silanes
and Shleifer (2008) and Djankov, Porta, Lopez-de-Silanes and Shleifer
(2003). Based on Tromp’s (2008) study, I generated a set of dummy
variables germane to FOIL features.
2. Methodology
The study employs four regression techniques as a main methodology to
analyze the panel data: 1) a simple OLS model without time dummy
variables; 2) an OLS model with time dummies; 3) a random effects
model (REM) or an error-components model (ECM) estimated by
generalized least square (GLS); and 4) a fixed effects model (FEM) to
consider time-specific and country-specific effects. The model
specifications are as follows.
For i-th country and t-th year,
R = realization of FOI ideals F = FOIL dummy (1 if a country has an FOIL)
Q = quality of public administration S = socio-political stability
L = legal tradition dummies T = year dummies
C = country dummies
The first model (OLS estimation in a pooled data) does not allow us to
know both the change in a dependent variable over time and the variation
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 113
in the dependent across countries at a given point in time. Estimates in
the second model, despite inclusion of time variables, are likely to be still
inflated and consistently biased due to inconsideration of hidden
cross-sectional effects. Thus, the best regression model for analyzing this
panel data is selected between FEM and REM.
Two panel data techniques have differential assumptions. While FEM
adds dummies (both Ci0 and T0t) to allow for the changes in the
cross-sectional and time-series intercepts, REM assumes error terms are
correlated across years and individual countries (Pindyck & Rubinfeld,
1997). REM treats an intercept term in FEM as two random components
(time-series error v0t and cross-sectional error ui0). The basic assumptions
underlying both models do not perfectly fit this panel data; reversely
saying, the dataset partially match both assumptions. If error terms and
regressors are correlated, FEM should be preferred over REM (Gujarati,
2002). Since I expect the size of error terms moves with the variation in
independents like the quality of public administration or political stability,
FEM would be more appropriate than REM. In the meantime, the
inclusion of binary variables (FOIL dummy and legal tradition dummies)
with a value constant over years substantially decreases an overall level of
the correlation between regressors and error terms. Accordingly, the
decline in the correlation between error terms and regressors weakens
imperativeness of selecting FEM. Moreover, the REM assumption is
tenable with the expectation that the error variance is correlated across
specific years and individual countries (heteroskedasticity).
Gujarati (2002) recommended that when the number of time-series data
is small but the number of cross-sectional units is large (the same
situation as in this dataset), REM is appropriate if the cross-sectional
units in the sample are regarded as random drawings from a much larger
population. However, the global dataset does not meet the REM
114 「정부행정」 제10권(2014)
assumption of random sampling because a group of available countries in
the whole sample is almost the same as the population. REM,
nevertheless, has a practical appeal from using up fewer degrees of
freedom as well as a conceptual appeal from broad characterization of the
sources of error (Pindyck & Rubinfeld, 1997). A practical problem of
FEM in analyzing this panel dataset is that it drops some of critical main
variables due to the collinearity between multiple dummies.
There is a simple rule for a choice between FEM and REM. The
Hausman test recommends we select FEM when we can reject the null
hypothesis that slope coefficients are equal between both models (Gujarati,
2002). Nonetheless, I cannot hinge on the convenient criterion without
consistency in both underlying assumptions. Since there is no last resort to
select one of two models, I consider both. Still, the panel data techniques
provide more meaningful implications and more precise (consistently
unbiased) estimates than an OLS model in a simply pooled data.
Along with this main methodology, I rely on ANOVA to examine the
significance of the mean difference in FOI achievement scores across
groups of countries (between countries with an FOIL and those without
the law, and among disparate legal traditions). Contingency tabulation and
charts will illustrate the cross-group difference.
3. Dependent variables: ATI, press freedom and transparency
Dependent variables represent three ideals of an FOIL discussed in the
previous section. The first variable is the level of citizens’ access to
government information. I employ five sub-indicators of Global Integrity
Index in 2004, 2006, 2007 and 2008. A newly created index is an average
of the following five percentage-scale scores: 1) “In practice, citizens
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 115
receive responses to access to information requests within a reasonable
time period”; 2) “In practice, citizens can use the access to information
mechanism at a reasonable cost”; 3) “In practice, citizens can resolve
appeals to access to information requests within a reasonable time
period”; 4) “In practice, citizens can resolve appeals to information
requests at a reasonable cost”; and 5) “In practice, the government gives
reasons for denying an information request.”
The second dependent variable is an index of press freedom provided
by Freedom House. It is an aggregate of three sub-indicators: 1) laws
and regulations that influence media contents; 2) political pressures and
controls on media contents; and 3) economic influences over media
contents. The summated score has a percentage scale. Islam (2006) found
that the presence of an ATI law is positively correlated with press
freedom and negatively correlated with journalist abuse. The indicator of
press freedom is expected to have a substantial correlation with whether a
country has an FOIL.
The third dependent variable is governmental transparency or openness,
which refers to the ability to find out what is going on inside government
(Piotrowski & Ryzin, 2007). I adopt Transparency International’s
Corruption Perception Index (CPI) as a proxy for the level of national
transparency. An individual country’s CPI score indicates the degree of
public sector corruption as perceived by business experts and government
analysts. An original value (interval scale) of the index ranges between 0
(highly corrupt) and 10 (highly clean and transparent). For visible
comparability with other dependent variables, I also use the
percentage-scale transformation (multiplied by 10) of the index. The
cross-national study by Relly and Sabharwal (2009) provides a ground for
adopting the three indicators as dependent variables, showing that
countries ranked as most transparent have significantly a higher level of
116 「정부행정」 제10권(2014)
mass media freedom and legal protection for ATI.
<Table 2> Descriptive statistics in 2007 data
Variable N of obs Mean Std. Dev. Min MaxPRESS FREEDOM 194 53.469 24.367 2 91TRANSPARENCY 180 3.993 2.090 1.401 9.423Access-to-information [ATI] 50 0.424 0.242 0 0.851*FOIL 216 0.333 0.472 0 1PUBLIC ADMINISTRATION 207 0 1 -2.374 2.188STABILITY (z-score) 209 0 1 -3.010 1.650*Legal tradition: ENGLISH 188 0.324 0.469 0 1*Legal tradition: FRENCH 188 0.532 0.500 0 1*Legal tradition: GERMAN 188 0.101 0.302 0 1*Legal tradition: SCANDINAVIAN 188 0.027 0.161 0 1*Legal tradition: SOCIALISM 188 0.016 0.126 0 1FOI-related variablesSpecificity 79 6.443 2.500 0 11History (the number of years) 72 11.708 12.027 0 61*Constitution 79 0.734 0.445 0 1*Who request 79 0.937 0.245 0 1*Anyone request 79 0.658 0.477 0 1*Citizen request 79 0.278 0.451 0 1*Fee 79 0.646 0.481 0 1*How to request 79 0.722 0.451 0 1*How to response 79 0.899 0.304 0 1*Public interest override 79 0.570 0.498 0 1*Officials’ duty 79 0.646 0.481 0 1*Officials’ penalty 79 0.494 0.503 0 1*Other laws override 79 0.810 0.395 0 1*Affirmative disclosure 79 0.759 0.430 0 1*Policy advice exemption 79 0.342 0.477 0 1*Cabinet record exemption 79 0.241 0.430 0 1*Whistleblower protection 79 0.139 0.348 0 1
*: binary variable (dummy)Note 1: Public administration and stability are measured in z-score. Note 2: Specificity of an FOIL is an aggregate of 11 binary items so that it has a minimum of 0 and a maximum of 11.
4. Independent variables: FOIL, administration, political stability and legal tradition
Including a dummy variable of whether a country has an FOIL is
important because its coefficient can signify the difference between FOIL
countries and no-FOIL countries. Also, this categorical variable is
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 117
necessary for finding out the relationship between the presence of the law
and the real level of FOI. Studies to date have regarded the presence of
the law as a dummy variable. Relly and Sabharwal (2009) examined
whether the presence of an ATI law influences business executives’
perceptions of national transparency in government policymaking. Using an
FOIL dummy, Islam’s (2006) analysis concluded the presence of an FOIL
affects governance. Unlike in other studies, the sample of FOIL countries
in this study includes countries – Argentina, China, Hong Kong, Pakistan
and Philippines – that have similar legal provisions as a form of code or
regulation though it is not exactly an FOIL. The longitudinal dataset
includes the latest FOIL joiners – Chile (2008), Cook Islands (2008),
Indonesia (2008), Jordan (2007), Nepal (2007), and Nicaragua (2007).
<Table 3> Collinearity between WGI indices
Variable Variation Influential Factor (VIF) Tolerance (1/VIF)WGI: Government Effectiveness 19.61 0.0510WGI: Rule of Law 16.80 0.0595WGI: Control for Corruption 14.01 0.0714WGI: Regulatory Quality 10.16 0.0984WGI: Political Stability 1.98 0.5051Mean VIF 12.53 –
<Table 4> Scoring by factor analysis
VariableFactor 1
(Eigenvalue = 3.71)Factor 2
(Eigenvalue = 0.03)Uniqueness
WGI: Government Effectiveness 0.9789 0.0623 0.0379WGI: Rule of Law 0.9583 -0.0811 0.0751WGI: Control for Corruption 0.9644 -0.0977 0.0603WGI: Regulatory Quality 0.9480 0.1170 0.0877
Focusing on the gap between legal ideals and realities, this study
hypothesizes factors influencing FOI realities are more significant than the
FOIL dummy. There are three influential factors. First, the general quality
of public administration will affect the level to which a country
118 「정부행정」 제10권(2014)
accomplishes the ideals. WGI provides four indicators related to public
administration. Table 3 describes substantial collinearities among its
sub-indicators. Four indicators (i.e., how effectively a national government
is operated, how well the rule of law is implemented, how well the public
sector controls corruption, and how well a regulation is implemented)
have too high correlation (low tolerance) to be included together in a
model (Table 3). They are combined into one score by factor scoring
(Table 4). I name the newly created score as Public Administration.
Contrary to other government-related WGI indicators, Political Stability
that has a lower VIF (less serious collinearity) is independently adopted as
another main explanatory variable to be a proxy for socio-political
stability in an individual country.
The third force to affect FOI ideals is legal tradition, which evolves
within a social context, and thus it reflects changes in society and culture.
This study employs the categorization of legal tradition used in the studies
by Djankov et al. (2003) and Porta et al. (2008). In their categorization,
all legal traditions fall into five heritages: English, French, German,
Scandinavian, and Socialist. Countries categorized into English heritage (the
United States and Commonwealth countries following Westminster
tradition) are regarded as to have common law tradition (Djankov et al.,
2003; Porta et al., 2008).
5. Independent variables: FOIL features
While an FOIL dummy may significantly differentiate between FOIL
countries and no-FOIL countries, multiple characteristics of an FOIL may
help explain significantly differences across countries within an FOIL
group. Thus, these variables are employed to answer the question of
“What FOIL features have an important influence on FOI ideals?” Across
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 119
countries, an FOIL has a common set of diverse provisions, but some
items bring fundamental differences in legal provisions. Except the history
of FOIL, I coded all FOIL features as a dichotomous status of yes (1) or
no (0).
FOIL history The history of an FOIL is to find out whether
countries with the law in presence for longer years are more likely to
raise the level of ATI, press freedom and governmental transparency. It is
measured as the number of years since the passage by subtracting the
adopted year (Appendix 1) from a specific year in the panel dataset. To
eliminate Sweden’s potential leverage as an outlier, I used the year of
revision (1949) instead of the precursor year of 1776 in global FOIL
history. Sweden, despite such transformation, is still the oldest FOIL
country in the dataset. Columbia that adopted a law on access as
prematurely as in 1885 is also suspected of another leverage as an outlier.
The initial law has been unused for a century until the year of 1985
when the modern law was established (Banisar, 2006). In my dataset, the
adopted year of Columbia is coded as 1985. Since most countries have
short history except some advanced countries, I use the first adopted year
in each FOIL country instead of a revision year.
Specificity Concreteness or specificity of an FOIL is measured by the
number of particular provisions (among 11 items that this study focuses
on) that it has. In addition to the ordinal variable, the study uses its
dichotomous transformation. If a country has more provisions than 6, the
binary variable is coded as 1 (more specific), and as 0 (less specific) for
otherwise.
Constitutional guarantee The constitutional status of FOI is a legal
precondition to protect specific statues of an FOIL constitutionally
(Tromp, 2008). If the national constitution states the public’s right to
know, this variable is coded as 1, and as 0 for otherwise.
120 「정부행정」 제10권(2014)
Who may ask (“Anyone can request,” “Only citizen can request”) A
majority of countries allow anyone to ask for information regardless of
citizenship, residence or interests (Banisar, 2006). Whereas there are
countries allowing for anonymous requests, some countries require a
requester to show proof of citizenship. This study uses two binary
variables: “anyone can request” and “only citizens can request.” Both are
mutually exclusive. If an FOIL has either one, the variable for the
provision of “who may ask” is coded as 1.
Fee Demanding fees from requesters limit the ability of the less well
off to ask for information. The premise “fees are barriers” existed as early
as in the time of Reagan administration (Feinberg, 1986). Governments’
attempts to sell information and increase FOI fees may create economic
barriers to openness (Roberts, 2000). With the aggressive pursuit of new
sources of nontax revenue, NPM governments in the United States have
attempted to package and sell information instead of releasing it at low
costs in response to FOI requests (Piotrowski & Rosenbloom, 2002). A fee
provision can reduce requests to a substantial degree so that a
government’s economic motive deters citizens and the press from exercising
their FOI rights. By imposing new fees on applications and appeals,
Ireland experienced the decline in the number of requests by half (Banisar,
2006). Commodification of government information may be a threat to
openness. The absence or presence of a fee provision is an important
dummy variable to predict the level of realizing FOI ideals.
How to request and How to response An FOIL in some countries
lacks description of procedures of request and response. Without
explanation of the step-by-step process from request to disclosure, the
FOIL may be no more than its title. Concreteness of the described
schedule will be correlated to performance of the law. If an FOIL
describes how to request and how to response, the variables would have a
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 121
value of 1.
Public interest override The provision of public interest override is
crucial for governmental transparency because it limits the scope of
exemptions. Public interest test requires public authorities to balance the
interest in withholding information against the public interest in disclosure
(Banisar, 2006). The provision contains the following items for public
interest test: 1) Should the provision apply to all statutory exemptions?;
2) Should it override exemptions absolutely? Or should other needs be
weighed and balanced against it?; and 3) Should the state have a duty to
proactively publish such information? (Tromp, 2008). If a country has a
clear provision about these items, this variable is coded as 1.
Duty to publish One of common principles in an FOI regime is an
obligation to publish; that is, public bodies should be under an obligation
to publish key information (Mendel, 2008). While some information is
routinely released, other information is proactively published under the
FOIL (Tromp, 2008). If an FOIL includes officials’ duty to publish
government-held information, the variable has a value of 1.
Noncompliance penalty The presence of a penalty provision is a vital
factor for greater openness (Tromp, 2008). A feature of FOILs designed to
encourage open administration is the sanctions provision to punish the
arbitrary and capricious withholding of documents (Gellman, 1997). Reylea
(1979) identified threatening penalties for the arbitrary or capricious
withholding of records as a key driver for faithful execution of an FOIL.
Without severe penalties, transparency requirements are merely aspirational
(BGA, 2002, 2008). The relevant provision stipulates penalties for officials
who destroy, falsify records, or delay, refuse responses to requests. If an
FOIL includes a penalty provision for officials’ noncompliance with the
law, the variable is coded as 1.
Other laws override Whether an FOIL allows other laws to override
122 「정부행정」 제10권(2014)
the statute on information release is a critical provision. Especially, a
secrecy provision related to national defense may override an FOI statute.
This variable is coded as 1 when an FOIL contains a provision that other
laws override it.
Affirmative disclosure An active provision of information publication
is a common feature of most FOILs (Banisar, 2006). Proactive disclosure
of government-held information implies voluntary, routine publication of
non-requested information. If a government releases particular information
without having to make an FOI request, the variable is coded as 1.
Policy advice exemption Unlike other FOIL exemptions of privacy
and national defense, an exemption for policy advice seems “innocuous”
(Tromp, 2008). This exemption is based on the assumption that the
actions and decisions of a government should not withstand too much
exposure. The public’s access to records on policy development might
hinder policymaking processes because “the threat of public scrutiny would
curb free and frank discussion, inhibit the candor of advice, and therefore
seriously hamper the smooth running of government” (Tromp, 2008). The
presence of the exemption for policy advice is coded as 1, no matter how
the provision is broad or narrow.
Cabinet records exemption Strict cabinet secrecy is a strong tradition
in Commonwealth countries (Tromp, 2008). Some specific types of cabinet
documents are excluded for cabinet confidences in FOIL applications. If an
FOIL has the provision of cabinet records exemption, the variable is
coded as 1.
Whistleblower protection This provision offers legal protection for
individuals who release information on wrongdoing (Mendel, 2008). It
would facilitate the disclosure of information on negative activities (corrupt
practices or mismanagement) in public agencies and give people an
incentive to come forward without fear of being sanctioned for their
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 123
disclosures (Banisar, 2006). If an FOIL contains a whistleblower protection
provision, this variable is coded as 1, otherwise as 0.
Ⅳ. Analysis
1. Between-group comparison
As illustrated in Figure 1, the mean difference between FOIL countries
and no-FOIL countries is significant in three dependent variables. FOIL
countries show a consistently higher level of citizens’ ATI, press freedom
and governmental transparency than no-FOIL countries. This trend
corresponds to the general tendency (significant χ2 statistics) in frequency
distribution in Table 5, but findings from conditional contingency in the
cross-tabulation are not as promising for FOI ideals as a simplistic mean
comparison in Figure 1 is.
<Figure 1> The mean difference in dependent variables between
FOIL countries and No-FOIL countries (% scale, 2007 data)
[ ] = Standard Deviation
124 「정부행정」 제10권(2014)
In Table 5, 29% of FOIL countries fall in high ATI group; whereas,
merely 6% of no-FOIL countries are categorized as high level of ATI.
The percentage distribution of FOIL vs. no-FOIL in the lowest level of
ATI (83%:17%) is opposite to that in the highest ATI group (18%:82%).
However, while 60% of FOIL countries lie in a middle level of ATI, the
equal proportion in the no-FOIL group takes the low level of ATI. The
percentage of high ATI countries in the FOIL group is not as great as
that of low ATI countries in the no-FOIL group. Thus, the significance in
the mean difference does not necessarily imply a sharp contrast of FOIL
vs. no-FOIL countries.
The distribution in the press freedom index is more supportive for the
effectiveness of an FOIL than that in the ATI index. While 55% of FOIL
countries have free press, only 27% of no-FOIL countries have free press.
The FOIL vs. no-FOIL contrast is also sharp in the not-free category,
but the frequency in the ideal (free) state is less desirable than that in the
non-ideal (not-free) state. In the not-free state of press freedom, the
ratio of FOIL to no-FOIL countries is 13%:87%; whereas, the ratio in
the free state is 54%:46%. The distribution in the ideal state does not
correspond to our high expectation for the positive impact of an FOIL on
press freedom.
Despite a general tendency of high transparency in FOIL countries, the
frequency in transparency is almost evenly distributed between the middle
(38%) and the high (40%) level. The contingency tabulation between
transparency and FOIL exhibits that in spite of the significant p-value in
the χ2 statistic, the middle vs. high contrast (38%:40%) in the FOIL
group is not as sharp as the frequency difference between low and high
transparency (46%:28%) in the same group. In no-FOIL countries, the
frequency is also equal between the middle (28%) and the high (26%)
level. Hence, t and χ2 statistics serve as a more lenient criterion to signify
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 125
the between-group difference than the comparison between cell
percentages in the cross-tabulation is.
<Table 5> Contingency between the FOIL dummy and dependent variables
(Row/Column/Cell %)χ2=42.30, pr=0.000
Practices of ATI (panel)
Low (0-33) Middle (34-66) High (67-100) Total
FOIL 10 (12/18/6) 50 (60/65/31) 24 (29/83/15) 84 (100/–/52)
No FOIL 46(59/82/28) 27 (35/35/17) 5 (6/17/3) 78 (100/–/48)
Total 56 (–/100/34) 77 (–/100/48) 29 (–/100/18) 162 (100/100/100)
(Row/Column/Cell %)χ2=25.51, pr=0.000
Press freedom (2007)
Not free Partly free Free Total
FOIL 8 (11/13/4) 24 (34/41/12) 39 (55/54/20) 71 (100/–/37)
No FOIL 55 (45/87/28) 35 (28/59/18) 33 (27/46/17) 123 (100/–/63)
Total 63 (–/100/32) 59 (–/100/30) 72 (–/100/37) 194 (100/100/100)
(Row/Column/Cell %)χ2=11.50, pr=0.003
Transparency (2007)
Low Middle High Total
FOIL 16(22/20/7) 27 (38/40/13) 29 (40/43/13) 72 (100/–/33)
No FOIL 66 (46/80/30) 40 (28/60/19) 38 (26/57/18) 144 (100/–/67)
Total 82 (–/100/37) 67 (–/100/32) 67 (–/100/31) 216 (100/100/100)
<Figure 2> The mean difference in dependent variables across
legal heritages (% scale, 2007 data)
[ ] = Standard DeviationNote: There is no ATI data of countries with Scandinavian and socialist legal heritage.
126 「정부행정」 제10권(2014)
On the other hand, Figure 2 portrays the mean difference in three
dependent variables across legal traditions. Overall, countries with
Scandinavian and German legal heritage have a higher level of indices.
English and French traditions have a within-group mean value around a
grand mean (total average). The between-tradition mean difference is
significant in all three variables. English and French groups have more
mean-droppers decreasing the average level or more outliers than
Scandinavian and German groups do. The distribution within English and
French groups, thus, reflects heterogeneity across countries with the same
legal tradition.
Charts and contingency tabulation sketch the visible distinction across
disparate groups – between FOIL countries and no-FOIL countries, and
among differential legal tradition groups. The hitherto analysis by
conditional frequency and mean comparison, however, does not bring a
valid evidence for the impact of an FOIL on achievement of its goals. I
posit determinants for FOI realities have a significant influence on
realization of FOI ideals. The next sections, presenting the result of
regression analyses on the panel data, test the significance of determinants
for estimating the level of ATI, press freedom and transparency.
2. Access to information
Table 6 describes the result of regressions for predicting the ATI level
for the whole sample. Whether a country has an FOIL is not significant
for estimating the ATI level when controlling for time-series and
cross-sectional effects. In the FEM that controls hidden effects from
individual countries and particular years, the presence of an FOIL does
not have an independent explanatory power. The impact of an FOIL
dummy on the ATI level is substantially diminished by direct or indirect
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 127
influences from year-specific and country-specific effects (influences by
omitted variables related to year or country).
<Table 6> Regressions of ATI for all countries
Simple OLS OLS with time FEM REM
FOIL 16.174† (3.778) 16.006† (3.783) 11.273 (16.157) 15.698† (4.013)
PUBLIC ADMIN 11.294** (3.318) 10.994** (3.333) 7.297(37.733) 11.229**(3.517)
STABILITY 0.970 (2.600) 0.782 (2.604) 10.022(15.817) 0.883(2.748)
ENGLISH -14.942* (6.713) -14.736* (6.718) Dropped -14.509* (7.105)
FRENCH -6.546 (6.420) -6.383 (6.424) Dropped -5.869 (6.780)
GERMAN Dropped Dropped Dropped Dropped
SCANDINAVIAN Dropped Dropped Dropped Dropped
year2006 – -3.526 (3.636) -3.135(4.269) -3.364(3.453)
Constant 77.322† (11.169) 78.371† (11.224) 69.445**(19.943) 76.064† (11.115)
N of obs 113 113 113 113
N of groups – – 71 71
Adjusted R2 0.4551 0.4548 0.3930 0.4886
F or χ2 F=16.59 (pr=0.00) F=14.35 (pr=0.00) F=0.87 (pr=0.51) χ2=89.24 (pr=0.00)
Hausman test – – χ2=1.27 (pr=0.94)
*: pr < 0.05, **: pr < 0.01, †: pr < 0.001Note: Dummy variables other than year2006, ENGLISH and FRENCH are dropped due to collinearity.
On the other hand, political stability is not a predictor for an individual
citizen’s ATI in all models. With the exception of FEM, the quality of
public administration is significant at the 1% level. Amongst legal
traditions, only English heritage is significant. The sign and magnitude of
the slope coefficient implies that as shown in an overall lower mean value
of an English heritage group in Figure 2, English tradition has a lower
score of ATI by 14.5 than German and Scandinavian tradition – here,
two dropped variables are a comparison category. The result of the
Hausman test supports preference of REM over FEM. The GLS estimation
of REM is almost the same as OLS estimation. Under the REM
assumption that an error term changes with respect to year and country,
the FOIL dummy, public administration and the English-tradition dummy
are a set of significant regressors to estimate ATI for all countries.
128 「정부행정」 제10권(2014)
The result of regression of ATI for the FOIL sample in Table 7 is
slightly different from Table 6. REM for the subsample still shows
significance of the quality of public administration, but its significance
drops to the 5% level. English tradition loses its significance. In this
model, public administration is solely a significant predictor for explaining
the level of citizens’ ATI. As in the regression of a whole sample, the
Hausman test suggests reliance on REM. However, this result lacks
efficiency of the regression estimates due to the small number of
observations and many degrees of freedom. Notwithstanding, it is
invariable that public administration is a more important variable than
others are.
<Table 7> Regressions of ATI for the FOIL subsample
Simple OLS OLS with time FEM REM
PUBLIC ADMIN 11.947* (4.633) 11.938* (4.706) 4.077(61.382) 12.315*(5.067)
STABILITY 0.336 (4.856) 0.332 (4.911) 1.891(19.872) -0.033(5.255)
ENGLISH -15.385 (8.890) -15.372 (9.020) Dropped -15.076 (9.772)
FRENCH -6.981 (7.462) -6.990 (7.553) Dropped -7.236 (8.155)
GERMAN Dropped Dropped Dropped Dropped
SCANDINAVIAN Dropped Dropped Dropped Dropped
year2006 – -0.104 (5.576) -1.964(6.568) -0.484(5.271)
Constant 84.362† (14.868) 84.398† (15.146) 77.624**(21.947) 83.981† (14.881)
N of obs 55 55 55 55
N of groups – – 34 34
Adjusted R2 0.2632 0.2479 0.2562 0.3310
F or χ2 F=4.86 (pr=0.001) F=3.97 (pr=0.003) F=0.41 (pr=0.800) χ2=21.13 (pr=0.002)
Hausman test – – χ2=1.27 (pr=0.94)
Note: Dummy variables other than year2006, ENGLISH and FRENCH are dropped due to collinearity.
3. Press freedom
Despite a general tendency of the similar result between OLS on the
pooled data and REM on the panel data, REM for estimating press
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 129
freedom drops a significance of the FOIL dummy. Though the Hausman
test recommends using FEM, its dropping of all legal traditions allows for
attention to REM. Legal heritage dummies are all significant in predicting
the level of press freedom. The magnitude is directly comparable. The
slope magnitude of Scandinavian and German traditions is much larger
than that of English and French traditions. The Scandinavian vs.
non-Scandinavian difference is as high as 64.6 scores.
The contrast between FEM and REM is notable in the variable of
public administration. When controlling for country-specific and
year-specific effects in FEM, the quality of public administration is not a
significant variable any more. However, with the REM assumption of
heteroskedasticity caused by year-variant and country-variant error terms,
public administration is highly significant in a modest positive magnitude.
Here, assumptions grounding both models are equally appealing to this
panel dataset. Thus, I underscore a common finding from two models.
Political stability is, in spite of the different level in significance, an
important predictor for estimating press freedom.
Regressions for the FOI subsample in Table 9 exhibit a dramatic change
in slope coefficients on legal tradition dummies. Not only OLS models but
also REM loses the significance level. Furthermore, the sign of English and
French dummy, despite insignificance of the coefficients, shifts from positive
to negative. This implies legal traditions do not significantly affect the level
of press freedom in the FOIL subsample. As like the result for the whole
sample, the Hausman test selects more appropriateness of FEM. In both
FEM and REM, political stability in countries with an FOIL is a significant
explanatory variable for predicting press freedom. The significance of public
administration fundamentally differentiates between FEM and REM.
Socio-political stability is persistently important in estimating press freedom
no matter how an individual country has an FOIL.
130 「정부행정」 제10권(2014)
<Table 8> Regressions of press freedom for all countries
Simple OLS OLS with time FEM REM
FOIL 9.853† (1.198) 9.853† (1.200) 0.129 (0.634) 0.591 (0.647)
PUBLIC ADMIN 9.068† (0.867) 9.068† (0.869) 1.709(0.923) 5.444† (0.835)
STABILITY 7.508† (0.778) 7.508† (0.779) 1.117*(0.505) 1.594** (0.509)
ENGLISH 35.478† (3.863) 35.478† (3.871) Dropped 45.921† (9.437)
FRENCH 31.364† (3.774) 31.364† (3.781) Dropped 37.221† (9.306)
GERMAN 32.932† (4.080) 32.932† (4.087) Dropped 51.403† (9.932)
SCANDINAVIAN 32.832† (5.024) 32.832† (5.033) Dropped 64.628† (11.844)
year2003 – 0.070 (1.500) -0.061(0.303) -0.070(0.316)
year2004 – -0.102 (1.422) -0.181(0.287) -0.158(0.300)
year2005 – 0.040 (1.419) -0.026(0.287) -0.019(0.300)
year2006 – 0.077 (1.487) 0.086(0.300) 0.087 (0.313)
Constant 21.068† (3.951) 21.074† (4.093) 54.784† (0.368) 13.361 (9.200)
N of obs 1,077 1,077 1,077 1,077
N of groups – – 181 181
Adjusted R2 0.6005 0.5990 0.5312 0.4793
F or χ2 F=203.20 (pr=0.00) F=134.97 (pr=0.00) F=5.50 (pr=0.00) χ2=211.81 (pr=0.00)
Hausman test – – χ2=80.20 (pr=0.00)
Note: Collinearity among dummies drops all legal tradition variables in FEM.
<Table 9> Regressions of press freedom for the FOIL subsample
Simple OLS OLS with time FEM REM
PUBLIC ADMIN 10.848† (0.912) 10.842† (0.918) -2.153(1.582) 5.572† (1.215)
STABILITY 8.023† (0.970) 8.035† (0.978) 5.871† (0.867) 6.337† (0.873)
ENGLISH 4.718* (2.039) 4.717* (2.051) Dropped -0.538 (3.467)
FRENCH 2.535 (2.116) 2.536 (2.128) Dropped -6.217 (3.256)
GERMAN 4.385* (2.105) 4.381* (2.117) Dropped Dropped
SCANDINAVIAN Dropped Dropped Dropped 4.436 (5.325)
year2003 – 0.176 (1.695) 0.549(0.408) 0.277(0.476)
year2004 – 0.179 (1.523) 0.326(0.381) 0.522(0.435)
year2005 – 0.382 (1.435) 0.232(0.378) 0.402(0.407)
year2006 – 0.360 (1.444) 0.086(0.300) 0.207 (0.405)
Constant 58.407† (2.441) 58.275† (2.798) 69.536† (0.975) 67.098† (2.741)
N of obs 349 349 349 349
N of groups – – 69 69
Adjusted R2 0.7912 0.7888 0.4558 0.7516
F or χ2 F=220.83 (pr=0.00) F=130.99 (pr=0.00) F=11.13 (pr=0.00) χ2=232.33 (pr=0.00)
Hausman test – – χ2=56.96 (pr=0.00)
Note: Collinearity among dummies drops all legal tradition variables in FEM.
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 131
4. Transparency
Regressions of transparency reveal an apparent difference between the
OLS models and the panel data models. The significance in both the
FOIL dummy and political stability distinguishes results by simple OLS
models from results by panel data techniques. Whether a country has an
FOIL is not a significant variable in the panel data regressions. While
political stability loses its significance in both panel data models, public
administration consistently has a significant influence on transparency.
Three legal tradition dummies are significant at 5% level, but their sign is
negative despite a moderate impact in magnitude of the slope coefficients.
Instead of considering a negative sign on the dummies, the coefficients are
interpreted as the relative distances in score of the proxy for transparency
across legal heritages because they make a difference in intercepts.
<Table 10> Regressions of transparency for all countries
Simple OLS OLS with time FEM REM
FOIL -0.159** (0.052) -0.160** (0.052) 0.104(0.074) 0.040(0.064)
PUBLIC ADMIN 2.153† (0.041) 2.156† (0.041) 0.591† (0.157) 1.912† (0.066)
STABILITY -0.109** (0.039) -0.111** (0.039) -0.059(0.073) -0.081(0.054)
ENGLISH -1.262† (0.212) -1.272† (0.212) Dropped -1.042* (0.431)
FRENCH -1.274† (0.207) -1.285† (0.207) Dropped -1.080* (0.423)
GERMAN -1.432† (0.217) -1.445† (0.217) Dropped -1.203** (0.447)
SCANDINAVIAN -0.226 (0.250) -0.240† (0.250) Dropped Dropped
year2003 – -0.103 (0.076) -0.073(0.040) -0.107*(0.042)
year2004 – -0.055 (0.068) -0.041(0.037) -0.068(0.038)
year2005 – 0.093 (0.065) 0.030(0.034) 0.056(0.036)
year2006 – 0.028 (0.065) 0.022 (0.034) 0.023(0.036)
Constant 5.267† (0.215) 5.357 (0.226) 4.160 (0.051) 5.184† (0.425)
N of obs 866 866 866 866
N of groups – – 177 177
Adjusted R2 0.9163 0.9165 0.8921 0.9147
F or χ2 F=1,184 (pr=0.00) F=792 (pr=0.00) F=2.65 (pr=0.007) χ2=2,009 (pr=0.00)
Hausman test – – χ2=101.79 (pr=0.00)
Note: The dummy of Scandinavian tradition is dropped due to collinearity in REM.
132 「정부행정」 제10권(2014)
With the rejection of the null hypothesis that FEM and REM are not
systematically different, the χ2 statistic by the Hausman test supports a
better fit of FEM in estimation of the transparency level. If the equivalent
significance of FEM and REM in model fitness allows us to accept
simultaneously the basic assumptions behind the two models, the quality
of public administration remains as a significant determinant for
governmental transparency.
The result of running the subsample is not different from that in a
whole sample. Public administration is significant across models. Legal
traditions are significant, but have a negative sign reflecting a relative
distance to an average of a comparison group. A noteworthy difference
between the whole sample and the subsample appears at the coefficient on
stability in REM. Its negative sign gains the significance in the subsample.
This does not indicate political stability negatively correlates with
transparency in countries with an FOIL. Table 12 reveals that high
pairwise correlation between political stability and transparency have
positive Pearson correlation coefficients in both the FOIL subsample
(r=0.74) and the whole sample (r=0.73).
Differently to reasonable coefficients in pairwise correlation, partial
correlation delivers results that may be counterintuitive. When controlling
other independent variables, the partial correlation coefficient of political
stability on transparency is negative but marginally small in its absolute
value. Moderately high pairwise correlation between transparency and
stability disappears in partial correlation due to the lockstep relation
among multiple variables included in regression. This mismatch between
partial correlation and pairwise correlation can be the compelling evidence
that public administration is a consistently important estimator for
predicting governmental transparency, but socio-political stability is not.
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 133
<Table 11> Regressions of transparency for the FOIL subsample
Simple OLS OLS with time FEM REM
PUBLIC ADMIN 2.356† (0.069) 2.359† (0.070) 0.866† (0.245) 2.197† (0.110)
STABILITY -0.251** (0.074) -0.258** (0.075) -0.194(0.130) -0.222*(0.104)
ENGLISH -0.709† (0.151) -0.710† (0.151) Dropped -0.929** (0.316)
FRENCH -0.940† (0.155) -0.943† (0.155) Dropped -1.208† (0.325)
GERMAN -1.181† (0.154) -1.182† (0.155) Dropped -1.325† (0.320)
SCANDINAVIAN Dropped Dropped Dropped Dropped
year2003 – -0.096 (0.123) -0.090(0.064) -0.128(0.068)
year2004 – -0.158 (0.112) -0.139*(0.060) -0.171**(0.062)
year2005 – 0.023 (0.107) -0.045(0.057) -0.038(0.060)
year2006 – -0.034 (0.107) -0.036 (0.056) -0.050(0.059)
Constant 4.686† (0.179) 4.793† (0.205) 4.821† (0.147) 5.128† (0.315)
N of obs 340 340 340 340
N of groups – – 68 68
Adjusted R2 0.9290 0.9287 0.9131 0.9297
F or χ2 F=740.19 (pr=0.00) F=442.65 (pr=0.00) F=2.86 (pr=0.007) χ2=1,082 (pr=0.00)
Hausman test – – χ2=39.59 (pr=0.00)
Note: The dummy of Scandinavian tradition is dropped due to collinearity.
<Table 12> Pairwise and partial correlation with transparency
Correlation PubicAdministration
PoliticalStability English French German Scandinavian
PairwiserY*Xi (All)
0.946(pr=0.000)
0.734(pr=0.000)
0.093(pr=0.002)
-0.327(pr=0.00)
0.141(pr=0.000)
0.436(pr=0.000)
Pairwise rY*Xi (FOIL)
0.955(pr=0.000)
0.743(pr=0.000)
0.166(pr=0.000)
-0.390(pr=0.000)
-0.058(pr=0.223)
0.503(pr=0.000)
Partial rY*Xi|X’s (All)
0.881(pr=0.00)
-0.073(pr=0.033)
-0.198(pr=0.00)
-0.205(pr=0.000)
-0.227(pr=0.000)
-0.037(pr=0.275)
Partial rY*Xi|X’s (FOIL)
0.881(pr=0.00)
-0.185(pr=0.001)
-0.250(pr=0.00)
-0.315(pr=0.000)
-0.386(pr=0.000) Dropped
The negative pairwise correlation is not necessarily counterintuitive. A
high level of political stability can have the tendency to reduce the
pressure for governmental openness, which is often motivated by the need
to keep an eye on a political opponent. In the United States history, FOIA
arose from conflict between the Democratic Congress and the Republican
president. The carrying-over Congress espoused the principle of openness
even when the Presidency shifted to the Democratic Party, much to
134 「정부행정」 제10권(2014)
President Johnson’s chagrin. The lesson penetrating through this historical
fact is that though a general tendency is that socio-political stability
enhances the level of FOI ideals, politicians and the public under instable
politics may actively and strongly push a government toward greater
openness.
5. FOIL features
Although most FOIL countries have followed a role model of an FOIL,
specific provisions vary with individual countries. Table 13 illustrates FOIL
features with the respect to a legal tradition. Scandinavian countries own
far longer history of an FOIL than other legal heritage groups do. The
specificity of an FOIL, measured by the number of provision items
considered in this study, significantly differentiates one tradition from the
others. FOILs with English tradition are more specific.
65% of French-tradition countries have a fee provision while a third
quarter of English and German heritage countries contain the provision in
their FOIL. Notably, no Scandinavian country has a fee provision.
Considering that fees may act as a barrier against FOI, the absence of a
fee provision can partly explain a consistently high level of Scandinavian
FOI scores. However, no fee-provision has two meanings: 1) like
Scandinavian countries, a national government does not impose any fee
(literally free request); or 2) the lack of the provision may be just an
evidence of a poorly drafted FOIL if fees are actually imposed to an
information requester in administrative practices.
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 135
<Table 13> FOIL features with the respect to a legal tradition
FOIL featuresLegal tradition
TotalEnglish French German Scandinavian
History (F=17.74)† 11.9 years 10.0 years 8.4 years 41.4 years 11.7 years
Specificity (F=9.15)† 8.5 items 5.3 items 6.7 items 5.6 items 6.5 items
Constitution (χ2=1.88) 75% 76% 88% 60% 73%
Who can request (χ2=6.46) 100% 85% 100% 100% 94%
How to request (χ2=3.85) 80% 65% 88% 60% 72%
How to response (χ2=4.78) 95% 82% 100% 80% 90%
Anyone request (χ2=2.70) 60% 62% 81% 80% 66%
Only citizen request (χ2=2.62) 40% 24% 19% 20% 28%
Fee (χ2=10.91)* 75% 65% 75% 0% 65%
Public interest override (χ2=13.05)** 85% 35% 63% 60% 57%
Duty (χ2=11.68)** 90% 47% 75% 80% 65%
Penalty (χ2=7.91)* 75% 47% 38% 20% 49%
Other laws override (χ2=9.36)* 90% 65% 94% 100% 81%
Affirmative disclosure (χ2=1.71) 85% 74% 75% 60% 76%
Policy advice exemption (χ2=17.84)† 70% 15% 31% 20% 34%
Cabinet record exemption (χ2=21.97)† 60% 9% 6% 40% 24%
Whistleblower protection (χ2=7.37)* 30% 15% 0% 0% 14%
*: pr < 0.05, **: pr < 0.01, †: pr < 0.001
Public interest override, officials’ duty, penalty for officials, other laws
override, policy advice exemption and cabinet record exemption show a
remarkable difference among the four legal traditions. While 85% of
English-tradition countries have a provision of public interest override,
only 35% of French-tradition countries have it. The proportion of
countries that have the provision to stipulate public officials’ duty to
publish government information is 90% in the English tradition, but
merely 47% in the French tradition is. This pattern also appears in the
penalty provision and the other-laws-override provision. The low
proportion in German and Scandinavian heritage countries has a penalty
provision. While almost all countries in other traditions have a provision
of other laws override, two thirds of French-tradition countries have it.
136 「정부행정」 제10권(2014)
Both policy advice exemption and cabinet record exemption are mostly
included in an FOIL of English-tradition countries. The other provisions
unmentioned have almost equal percentage across traditions to make little
significance in the frequency tabulation (low χ2 statistics).
<Table 14> The mean values of dependent variables in terms of FOIL features
FOILfeatures
ATIM=53.12 (s.d=20.89)
Press freedomM=64.23 (s.d=20.71)
TransparencyM=4.96 (s.d=2.32)
SigGroup
SigGroup
SigGroup
YES NO YES NO YES NO
Specificity F=0.78 54.53 50.63 F=0.67 62.51 66.41 F=0.23 4.85 5.11
Constitution F=0.02 53.29 52.53 F=0.33 65.02 61.84 F=0.95 5.11 4.50
Anyone requests F=6.18** 57.52 47.12 F=23.24† 71.17 49.80 F=10.16† 5.53 3.82
Only citizen requests F=1.55 49.69 55.15 F=19.20† 48.55 69.74 F=5.70** 3.94 5.34
Fee F=2.68 55.10 47.13 F=8.62† 59.38 73.22 F=3.31* 4.61 5.62
How to request F=0.11 52.73 54.45 F=0.34 63.39 66.48 F=1.30 4.78 5.47
How to response F=3.38* 54.02 36.60 F=0.06 64.04 65.88 F=0.14 5.00 4.68
Public interest F=2.58 55.66 48.63 F=0.00 64.14 64.35 F=0.16 5.06 4.84
Officials' duty F=3.38* 56.35 48.53 F=3.27* 67.24 58.35 F=4.47** 5.37 4.20
Officials' penalty F=1.47 51.32 56.78 F=4.79** 59.26 69.34 F=10.64† 4.15 5.80
Other laws override F=15.80† 57.37 38.64 F=12.54** 68.06 48.40 F=5.66** 5.27 3.73
Affirmative F=0.37 52.53 55.94 F=0.29 63.50 66.47 F=1.40 4.79 5.53
Policy exemption F=4.87** 59.29 49.68 F=0.91 67.48 62.67 F=1.67 5.47 4.72
Cabinet records F=0.91 48.56 54.02 F=4.40** 73.00 61.56 F=1.43 5.53 4.78
Whistleblower F=3.10* 44.81 54.77 F=2.40 55.36 65.71 F=3.20* 3.76 5.15
*: pr < 0.10, **: pr < 0.05, †: pr < 0.01Note: An individual country belongs to the YES group when its FOIL has a specific provision.
FOIL features do not only present a distinction among legal traditions
but also play a role as determinants for the degree of fulfilling FOI ideals.
In Table 14, anyone request, officers’ duty and other laws override are an
FOIL’s important property to discriminate between the high and the low
level of achieving FOI ideals. The presence of a provision to guarantee
anyone’s request significantly increases scores in ATI, press freedom and
transparency. The index score pattern in the provision to limit a request
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 137
to a person with legal citizenship is apparently opposite to that in the
anyone-request provision. Countries where only citizens can request
information have low scores in the indicators of FOI ideals. The absence
of a fee provision enhances press freedom and transparency, but the
provision is not significant in making a difference in the ATI level. How
to respond is significant only in the mean difference of the ATI level.
Informing citizens of FOIL procedures can help enhance the level of access
to government information. Countries with a provision to describe
officials’ duty to publish exhibit higher scores in all FOI ideals than a
group of countries without the provision.
On the other hand, the mean difference in a penalty provision violates
the common-sensical expectation that the stipulation of penalty for
officials would increase the degree of fulfilling FOI ideals. Countries that
do not have the provision in their FOIL are persistently higher in all
dependent variables than countries with it. When an FOIL has a
whistleblower protection, its score for FOI ideals is significantly lower.
These counter-expected findings can attribute to a cross-cultural
discrepancy within a country. Both noncompliance penalties and
whistle-blower protection may only be of value in heterogeneous and
conflicting cultures rather than in homogeneous cultures as in Scandinavian
countries. In addition, the efficacy of an FOIL can be explained by
citizens’ expectation for public officials better than by the complete
presence of relevant legal provisions such as penalty and whistleblower
protection. Citizens in cultures of high FOI ideals tend to expect that
public officials would act not as independent agents but rather as agents
of the electorate. Such expectation enhances the perceived level in
achievement of FOI ideals in a country.
Ironically, the presence of some provisions to limit a scope of an FOIL
displays higher scores in FOI ideals than their absence does. When other
138 「정부행정」 제10권(2014)
laws can legitimately override an FOIL, scores of FOI ideals are
significantly higher than otherwise. This counter-intuitive pattern also
appears in policy advice exemption and cabinet record exemption.
Differently from the expectation that an exemption provision dampens FOI
ideals, ATI score is higher in a group of countries with a
policy-advice-exemption provision than in countries without the provision.
Likewise, countries with a provision of cabinet record exemption exhibited
higher score of press freedom than countries without the provision.
In the face of significance in several FOI features, the sign of its mean
difference brings an unexpected result. This result may come from an
obvious contrast of German and Scandinavian vs. English and French
tradition. Scores in all three dependent variables are high in German and
Scandinavian legal traditions. Described in Table 13, the two traditions
with overall high scores have FOIL features quite different from English
and French tradition with relatively low scores. For example, the presence
of a whistleblower protection and a penalty provision is expected to
increase scores of FOI ideals; however, German- and
Scandinavian-tradition countries with high mean scores do not have the
provisions in their FOIL. The mean difference between two tradition pairs
(German and Scandinavian vs. English and French) in FOI scores makes a
counter-intuitive result in terms of the absence or presence of a specific
provision.
Ⅴ. Discussions
The aforementioned section explored ideals and realities of an FOIL
through statistical analysis. The simple comparison of FOIL vs. no-FOIL
countries demonstrated a group of countries with an FOIL has
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 139
accomplished FOI ideals more than no-FOIL countries. However, the
regression analysis on the panel data overturns the monolithic finding from
the comparison by the mean difference. Whether a country has an FOIL
is not per se a promising indicator for realization of FOI ideals. The
existence of an FOIL may be itself rhetoric, and thus it is not any
guarantee for the ideal reality. Instead, this study confirms the presence of
important determinants for FOI ideals. While the FOIL dummy has little
significance in the panel data models considering influences from years and
countries, the quality of public administration and political stability are
significant predictors for estimating the level of FOI ideals.
On the other hand, legal traditions significantly distinguish high
achievers of FOI ideals from low achievers of them. Countries belonging
to German and Scandinavian tradition are more likely to accomplish FOI
ideals than those with English and French tradition within which the
range of FOI indicator scores is much wider due to cross-national
heterogeneity in diverse contexts of society, culture, and politics.
The final focus was casted on the relationship of statute features with
FOI ideals. Given the intention behind particular provisions, some of
significant differences in FOI ideals scores between countries with a
specific provision and those without it seem to break conventional
wisdom. It is unexpected that countries with exemption provisions have
higher scores of FOI. This counter-intuitive result hints that the degree of
specificity in an FOIL is important. The content of a provision can itself
be consequential, but this study found that concrete and sophisticated
stipulation gains an importance over the intention of a provision. This
finding accords with Gellman’s (1997) argument that one of problems
inherited in the FOIA may come from its poor draft.
140 「정부행정」 제10권(2014)
Ⅵ. Limitations
Concentrating on methodological concerns, I confess three limitations of
this study. First, it is no doubt that public administration and political
stability have a significant impact on realization of FOI ideals, but this
paper does not consider the independent impact from culture of secrecy in
the public sector and over a whole society. The prevalence of such
culture, to a substantial extent, predetermines the degree to which a
government accomplishes FOI ideals (Banisar, 2006). Considering culture
of secrecy in an empirical study would face methodological difficulties in
measuring culture quantitatively.
Another issue is inherent in a cross-national study whose unit of
analysis is an individual country. If implementation of an FOIL is
problematic, an individual agency is appropriate as a unit of analysis
because the lack of resources such as staff and budget and bureaucratic
attitude unfavorable for FOIL implementation are principal constraints
with an FOIL. However, a researcher confronts difficulty and even
impossibility to survey individual agencies across many countries. For a
cross-national analysis, this study employs a country’s quality of public
administration as if it can reliably represent a single level to aggregate all
governments in each country. This cross-national study assumes that the
general quality of public administration in a country significantly influences
the degree to which the stated goals of its FOIL are achieved. The idea is
not a deviation from the existing insightful albeit non-empirical argument
that administrative discretion and agency culture affect the realization of
the legal ideals. Rather, my assumption is still consistent with the
agency-level arguments because this study extends the scope of public
administration from an individual agency to a national government.
The third issue involves a statistical concern. Although panel data
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 141
techniques provide more implicative findings and more accurate estimation
than OLS on the pooled data does, the study lays little attention to high
likelihood of interpenetrations among an FOIL dummy, legal traditions
and FOI reality determinants. Including the interactive variables causes a
problem of soaking up many degrees of freedom, however. A statistically
better analysis, therefore, requires the increase in the number of
observations. Since the number of countries is fixed, the time-series
increase is the way to enlarge the size of the global longitudinal data.
With greater precision of statistical estimation and better-fitting model
specification including interaction terms, an analysis on the larger dataset
in the future needs to revisit implications from this study.
Ⅶ. Conclusions
A freedom-of-information law seeks for making a government
accountable for and transparent to the public by enhancing access to
information. The pursuit of FOI ideals is being realized in some countries,
but other countries use the presence of the law as a symbol of FOI.
Though FOIL countries accomplish FOI ideals more than no-FOIL
countries, the regression analysis on the panel data revealed the skeptic
finding that an FOIL itself is not a promoter to improve access to
information, press freedom and governmental transparency. Both the
quality of public administration and political stability obtain a greater
significance than whether or not a country has an FOIL.
By adding legal heritage dummy variables to regression, I found that
countries with German and Scandinavian traditions are overall reaching
closer to FOI ideals than followers of English and French traditions.
However, this is no more than a general tendency on an average statistic.
142 「정부행정」 제10권(2014)
The latter heritages have much broader variation in the range of FOI
scores than the former. Socio-cultural heterogeneity across countries within
a larger group of the same legal heritage contributes to the wider range
for estimating the degree to which an individual country accomplishes FOI
ideals.
Some FOIL provisions distinguish high achievers from low achievers of
FOI ideals. Countries that have provisions for anyone request, duty to
publish, noncompliance penalty and other laws override are more likely to
have high scores in FOI ideals indices. However, there is a
counter-intuitive but impressive finding; FOI ideals are more accomplished
in countries that have detailed exemption provisions than in countries
without such provisions. The punch line is that not only the content but
also concreteness of an FOIL makes the law come to the reality beyond
its rhetoric.
References
Banisar, D. (2006). Freedom of Information Around the World 2006: Global Survey of Access to Government Information Laws. London, U.K.: Privacy International, from http://www.privacyinternational.org/foi/foisurvey2006.pdf
BGA. (2002). The BGA Integrity Index. Chicago, IL: The Better Government Association, from http://www.bettergov.org/IntegrityIndexRelease.aspx.
BGA. (2008). The BGA - Alper Integrity Index. Chicago, IL: The Better Government Association, from http://www.bettergov.org/IntegrityIndexRelease.aspx.
Chongkittavorn, K. (2002). The media and access to information in Thailand. In The Right to Tell: The Role of Mass Media in Economic Development (pp. 255-265). Washington, D.C.: World Bank.
Darch, C., & Underwood, P. G. (2005). Freedom-of-information legislation, state compliance and the disclosure of knowledge: The South African experience. The International Information & Library Review, 37, 77-86.
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 143
Davis, C. N. (2000). A dangerous precedent: The influence of Critical Mass III on exemption 4 of the federal Freedom-of-information Act. Communication Law & Policy, 5(2), 183-202.
Djankov, S., Porta, R. L., Lopez-de-Silanes, F., & Shleifer, A. (2003). Courts. Quarterly Journal of Economics, 118(May), 453-517.
Feinberg, L. E. (1986). Managing the Freedom of Information Act and federal information policy. Public Administration Review, 46(6), 615-621.
Feinberg, L. E. (2004). FOIA, federal information policy, and information availability in a post-9/11 world. Government Information Quarterly, 21(4), 439-460.
Foerstel, H. N. (1999). Freedom-of-information and the Right to Know: The Orgins and Application of the Freedom-of-Information Act. Westport, CT: Greenwood Press.
Gellman, R. (1997). Shortcomings of the proposed Electronic Freedom of Information Improvement Act (S.1090). Government Information Quarterly, 14(1), 1-10.
Gujarati, D. N. (2002). Basic Econometrics. New York: Mcgraw-Hill.
Halstuck, M. E. (2000). Speed bumps on the information highway: A study of federal agency compliance with the Electronic Freedom-of-information Act of 1996. Communication Law & Policy, 5(4), 423-468.
Islam, R. (2006). Does more transparency go along with better governance? . Economics & Politics, 18(2), 121-167.
Lombard, D. R. (2007). Curiosity Killed the Cat: A Report on Compliance with Illinois's Freedom of Information Act. Chicago, IL: The Better Government Association, from http://www.bettergov.org/pdfs/BGA_FOIA_report_full.pdf.
Martin, R., & Feldman, E. (1998). Access to Information in Developing Countries. Berlin, Germany: Transparency International, from http://www.transparency.org/working_papers/martin-feldman/mart-feld.zip.
Mendel, T. (2008). Freedom of Information: A Comparative Legal Survey (2nd ed.). Paris, France: the United Nations Scientific and Cultural Organization (UNESCO), from http://unesdoc.unesco.org/images/0015/001584/158450e.pdf.
Meyer, J., & Rowan, B. (1991). Institutionalized organizations: Formal structure as myth and ceremony. In W. W. Powell & P. J. DiMaggio (Eds.), The New Institutionalism in Organizational Analysis (pp. 41-62). Chicago, IL: University of Chicago Press.
Mistry, J. (2006). Freedom of Information: Promoting Good Governance in Sierra Leone. New Delhi, India: the Commonwealth Human Rights Initiative, from
144 「정부행정」 제10권(2014)
http://www.humanrightsinitiative.org/new/2006/foi_promoting_good_govrn_sierraleonne.pdf.
North, D. C. (1999). Institutions, Institutional Change and Economic Performance. Cambridge, U.K.: Cambridge University Press.
Perritt, H. H., Jr., & Lhulier, C. J. (1997). Information access rights based on international human rights law. Buffalo Law Review, 45(Fall), 899-929.
Pindyck, R. S., & Rubinfeld, D. L. (1997). Econometric Models and Economic Forecasts (4th ed.). Boston, MA: McGraw-Hill.
Piotrowski, S. J., & Rosenbloom, D. H. (2002). Nonmission-based values in results-oriented public management: the case of freedom of information. Public Administration Review, 62(6), 643-657.
Piotrowski, S. J., & Ryzin, G. G. V. (2007). Citizen attitudes toward transparency in local government. The American Review of Public Administration, 37(3), 306-323.
Porta, R. L., Lopez-de-Silanes, F., & Shleifer, A. (2008). The economics consequences of legal origins. Journal of Economic Literature, 46`(2), 285-332.
Relly, J. E., & Sabharwal, M. (2009). Perceptions of transparency of government policymaking: A cross-national study. Government Information Quarterly, 26(1), 148-157.
Relyea, H. C. (1979). Faithful execution of the FOI Act: A legislative branch perspective. Public Administration Review, 39, 328-332.
Relyea, H. C. (2009). Federal freedom of information policy: Highlights of recent developments. Government Information Quarterly, 26(2), 314-320.
Roberts, A. S. (2000). Less government, more secrecy: Reinvention and the weakening of freedom of information law. Public Administration Review, 60(4), 308-320.
Roberts, A. S. (2005). Spin control and FOI: Lessons for the UK from Canada. Public Administration, 83(1), 1-23.
Rush, T. (1986). Journalists and the Freedom of Information Act. Public Administration Review, 46(6), 612-614.
Tromp, S. L. (2008). Fallen Behind: Canada's Access to Information Act in the World Context. from http://www3.telus.net/index100/foi.
Worthy, B. (2008). The future of freedom of information in the United Kingdom. The Political Quarterly, 79(1), 100-108.
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 145
Appendix 1. The list of countries with an FOI law
Country Year passed The formal name of an FOI Law
Albania 1999 The Law on the Right to Information for Official Documents*
Angola 2002 The Law on Access to Documents Held by Public Authorities*
Antigua and Barbuda 2004 Freedom of Information Act 2004
Argentina 2003 Access to Public Information Regulation†
Armenia 2003 The Law on Freedom of Information
Australia 1982 Freedom of Information Act 1982
Austria 1987 Federal Law on the Duty to Furnish Information*
Belgium 1994 The Law on the Right of Access to Administrative Documents*
Belize 1994 Freedom of Information Act
Bosnia and Herzegovina 2001 Freedom of Access to Information Act (FoAIA)
Bulgaria 2000 Access to Public Information Act
Canada 1982 Access to Information Act
Chile 2008 Freedom of Information Act
China 2007Ordinance on Openness of Government Information (OGI regulation)†
Columbia 1985 The Law Ordering the Publicity of Official Acts and Documents*
Cook Islands 2008 Freedom of Information Act
Croatia 2003 The Act on the Right of Access to Information
Czech Republic 1999 The Law on Free Access to Information
Denmark 1970 Access to Public Administration Files Act
Dominican Republic 2004 The Law on Access to Information
Ecuador 2004 Organic Law on Transparency and Access to Public Information
Estonia 2000 Public Information Act
Finland 1951 The Act on the Openness of Government Activities
France 1978 The Law on Access to Administrative Documents
Georgia 1999General Administrative Code of Georgia (Ch 3. Freedom of Information)
Germany 2005 The Act to Regulate Access to Federal Government Information
Greece 1986 The Code of Administrative Procedure
Honduras 2006 Access to Information Law
Hong Kong 1995 The Code on Access to Information†
Hungary 1992Protection of Personal Data and Disclosure of Data of Public Interest
Iceland 1996 Information Act
India 2005 Right to Information Act
Indonesia 2008 The Law Regarding Transparency of Public Information
Ireland 1997 Freedom of Information Act
Israel 1998 Freedom of Information Law
Italy 1990 Law No. 241
Jamaica 2002 Access to Information Act
Japan 1999The Law Concerning Access to Information Held by Administrative Organs
Jordan 2007 The Law on Securing the Right to Information Access
Korea, South 1996 The Act on Disclosure of Information by Public Agencies
Kosovo 2003 The Law on Access to Official Documents
146 「정부행정」 제10권(2014)
Source: Banisar, D. (2006); Tromp, S. L. (2008)†: The provision (code or regulation) is not the FOI law itself, but plays a part of similar functions. *: The country does not provide the formal name of the law in English.
Kyrgyzstan 2006 The Law on Access to Information
Latvia 1998 The Law on Freedom of Information
Liechtenstein 1999 The Information Act
Lithuania 1999 The Law on the Provision of Information to the Public
Macedonia 2006 The Law on Free Access to Information of Public Character
Mexico 2002The Federal Law of Transparency and Access to Public Government Information
Moldova 2000 The Law on Access to Information
Montenegro 2005 The Law on Free Access to Information
Nepal 2007 Right to Information Act 2007
Netherlands 1978 Government Information (Public Access) Act
New Zealand 1982 Official Information Act
Nicaragua 2007 The Law on Access to Information*
Norway 1970 Freedom of Information Act
Pakistan 2002Ordinance to provide for transparency and freedom of information†
Panama 2001 The Law on Transparency in Public Administration*
Peru 2002 The Law of Transparency and Access to Public Information
Philippines 1987Code of Conduct and Ethical Standards for Public Officials and Employees†
Poland 2001 The Law on Access to Public Information
Portugal 1993 The Law of Access to Administrative Documents (LADA)
Romania 2001 The Law Regarding Free Access to Information of Public Interest
Serbia 2004 The Law on Free Access to Information of Public Importance
Slovakia 2000 The Act on Free Access to Information
Slovenia 2003 Access to Public Information Act (ZDIJZ)
South Africa 2000 Promotion of Access to Information Act (PAIA)
Spain 1992 Law on Rules for Public Administration*
St. Vincent and the Grenadines
2003 Freedom of Information Act
Sweden 1949 Freedom of the Press Act
Switzerland 2004 Federal Law on the Principle of Administrative Transparency
Tajikistan 2002 The Law of the Republic of Tajikistan on Information
Thailand 1997 Official Information Act
Trinidad and Tobago 1999 Freedom of Information Act
Turkey 2003 The Law on Right to Information
Uganda 2005 Access to Information Act 2005
Ukraine 1992 The Law on Information
United Kingdom 2000 Freedom of Information Act 2000
United States 1996 Freedom of Information Act
Uzbekistan 2002The Law on the Principles and Guarantees of Freedom of Information
Zimbabwe 2002 Access to Information and Privacy Protection Act (AIPPA)
The Rhetoric and the Reality of the Legal Guarantee for Freedom of Information 147
Appendix 2. Categorization of countries in terms of legal tradition
Legal traditions
Countries (188)
No-FOI law countries (115) FOI law countries (73)
English(61)
Anguilla, Bahamas, Bahrain, Bangladesh, Barbados, Bermuda, Bhutan, Botswana, Cayman Islands, Cyprus, Dominica, Fiji, Gambia, Ghana, Grenada, Guyana, Kenya, Lesotho, Liberia, Malawi, Malaysia, Maldives, Micronesia, Namibia, Nigeria, Papua New Guinea, Samoa, Saudi Arabia, Sierra Leone, Singapore, Solomon Islands, Sri Lanka, St. Kitts and Nevis, St. Lucia, Sudan, Swaziland, Tanzania, Tonga, United Arab Emirates, Vanuatu, Zambia (41)
Antigua and Barbuda, Australia, Belize, Canada, Hong Kong, India, Ireland, Israel, Jamaica, Nepal, New Zealand, Pakistan, South Africa, St. Vincent and the Grenadines, Thailand, Trinidad and Tobago, Uganda, United Kingdom, United States, Zimbabwe (20)
French(100)
Afghanistan, Algeria, Aruba, Azerbaijan, Benin, Bolivia, Brazil, Burundi, Cambodia, Cameroon, Cape Verde, Central African Republic, Chile, Comoros, Congo Democratic Republic, Congo Republic, Costa Rica, Cote D’Ivoire, Djibouti, Egypt, El Salvador, Equatorial Guinea, Eritrea, Ethiopia, Gabon, Guatemala, Guinea, Guinea-Bissau, Haiti, Indonesia, Iran, Iraq, Kazakhstan, Kuwait, Laos, Lebanon, Luxembourg, Macao, Madagascar, Mali, Malta, Mauritania, Mauritius, Morocco, Mozambique, Netherlands Antilles, Niger, Oman, Paraguay, Puerto Rico, Qatar, Russia, Rwanda, Sao Tome and Principe, Senegal, Seychelles, Suriname, Syria, Togo, Tunisia, Turkmenistan, Uruguay, Venezuela, Vietnam, Yemen (68)
Albania, Angola, Argentina, Belgium, Colombia, Dominican Republic, Ecuador, Estonia, France, Greece, Honduras, Italy, Jordan, Kyrgyzstan, Lithuania, Macedonia, Mexico, Moldova, Montenegro, Netherlands, Nicaragua, Panama, Peru, Philippines, Portugal, Romania, Serbia, Spain, Tajikistan, Turkey, Ukraine, Uzbekistan (32)
German(19)
Belarus, Mongolia, Taiwan (3) Austria, Bosnia-Herzegovina, Bulgaria, China, Croatia, Czech Republic, Georgia, Germany, Hungary, Japan, Korea South, Latvia, Poland, Slovakia, Slovenia, Switzerland, (16)
Scandinavian(5)
Denmark, Finland, Iceland, Norway, Sweden (5)
Socialism (3)
Cuba, Korea North, Myanmar (3)
148 「정부행정」 제10권(2014)
Appendix 3. The list of countries included in Global Integrity Index
(ATI variables)
Year Countries
2004(25)
Argentina, Australia, Brazil, Germany, Ghana, Guatemala, India, Indonesia, Italy, Japan, Kenya, Mexico, Namibia, Nicaragua, Nigeria, Panama, Philippines, Portugal, Russia, South Africa, Turkey, Ukraine, United States, Venezuela, Zimbabwe
2006(41)
Argentina, Armenia, Azerbaijan, Benin, Brazil, Bulgaria, Congo Democratic Republic, Egypt, Ethiopia, Georgia, Ghana, Guatemala, India, Indonesia, Israel, Kenya, Kyrgyzstan, Lebanon, Liberia, Mexico, Montenegro, Mozambique, Nepal, Nicaragua, Nigeria, Pakistan, Philippines, Romania, Russia, Senegal, Serbia, Sierra Leone, South Africa, Sudan, Tajikistan, Tanzania, Uganda, United States, Vietnam, Yemen, Zimbabwe
2007(50)
Algeria, Argentina, Armenia, Azerbaijan, Bangladesh, Bosnia-Herzegovina, Bulgaria, Burundi, Cameroon, Canada, China, Colombia, Costa Rica, Ecuador, Egypt, France, Georgia, India, Italy, Japan, Jordan, Kazakhstan, Kenya, Kyrgyzstan, Latvia, Lebanon, Malawi, Mexico, Moldova, Mozambique, Namibia, Nepal, Nigeria, Pakistan, Peru, Philippines, Romania, Russia, Sierra Leone, Spain, Sri Lanka, Tajikistan, Tanzania, Thailand, Timor-Leste, Turkey, Uganda, Ukraine, United States, Vanuatu
2008(46)
Angola, Argentina, Azerbaijan, Bangladesh, Belarus, Bosnia-Herzegovina, Bulgaria, Cameroon, Canada, Chile, China, Colombia, Ecuador, Egypt, Ethiopia, Georgia, Ghana, Hungary, Indonesia, Iraq, Italy, Japan, Kazakhstan, Kenya, Kyrgyzstan, Lithuania, Macedonia, Moldova, Montenegro, Morocco, Nepal, Nicaragua, Nigeria, Pakistan, Philippines, Poland, Romania, Russia, Serbia, Somalia, South Africa, Turkey, Uganda, West Bank Gaza, Yemen
Paneldata(84)
Algeria, Angola, Argentina, Armenia, Azerbaijan, Australia, Bangladesh, Belarus, Benin, Brazil, Bosnia-Herzegovina, Bulgaria, Burundi, Cameroon, Canada, Chile, China, Colombia, Congo Democratic Republic, Costa Rica, Ecuador, Egypt, Ethiopia, France, Georgia, Germany, Ghana, Guatemala, Hungary, India, Indonesia, Iraq, Israel, Italy, Japan, Jordan, Kazakhstan, Kenya, Kyrgyzstan, Latvia, Lebanon, Liberia, Lithuania, Macedonia, Malawi, Mexico, Moldova, Montenegro, Morocco, Mozambique, Namibia, Nepal, Nicaragua, Nigeria, Panama, Pakistan, Peru, Philippines, Poland, Portugal, Romania, Russia, Senegal, Serbia, Sierra Leone, Somalia, South Africa, Spain, Sri Lanka, Sudan, Tajikistan, Tanzania, Thailand, Timor-Leste, Turkey, Uganda, Ukraine, United States, Vanuatu, Venezuela, Vietnam, West Bank Gaza, Yemen, Zimbabwe